Consider this line
[Title "Ponziani"]
[fen ""]
[Startply "14"]
1. e4 e5 2. Nf3 Nc6 3. c3 Nf6 4. d4 Nxe4 5. d5 Bc5 6. dxc6 Bxf2 7. Ke2 Bb6 8. Qd5 (8. Qc2) Nf2 9. Rg1 bxc6
I saw here that Eric Rosen played 8. Qd5.
This now allows 8. ..Nf2 9. Rg1 bxc6, opening the light-squared bishop with tempo.
Wouldn't 8 Qc2 be an improvement? The queen is safe there, the knight is still hit, and after the rook moves, I don't see a good way to stop Be3.
Did I miss something?
After 9...bxc6, white has 10.Qxe5+. If 10...Qe7, then white can exchange off the queens and a piece with 11.Qxe7+ Kxe7 12.Be3. So black plays 10...Kf8, and then white can block the bishop's path with 11.Nd4, and black's knight is in some trouble. Here black doesn't get enough compensation for the sacrificed piece.
So instead of 9...bxc6, black usually replies 9...0-0, sacrificing a second piece after 10.cxb7 Bxb7 11.Qxb7. Now the e-pawn is poisoned, white can't allow the e-file to open. And sometimes black can get a rook into play quickly with Ra8-b8xb2+. Amazingly, engines say that black has compensation for the two pieces here and may even be slightly better -- but in practice it's really hard to calculate everything for both sides, in my experience.
The main line ends up in a repetition draw though:
[FEN ""]
1.e4 e5 2.Nf3 Nc6 3.c3 Nf6 4.d4 Nxe4 5.d5 Bc5 6.dxc6 Bxf2+ 7.Ke2 Bb6 8.Qd5 Nf2 9.Rg1 O-O 10.cxb7 Bxb7 11.Qxb7 Qf6 12.Qd5 c6 13.Qd2 e4 14.Nd4 e3 15.Qc2 Bxd4 16.cxd4 Ng4 17.Rh1 Nf2 18.Rg1 Ng4 {=}
